Abstract: Vehicles, which operate more or less autonomously, are included in the large family of robots. The autonomy is tightly related with the use of sensors for environment perception. The mobile robots are actors on which there can be implemented strategies that belong to applied artificial intelligence.
A little wheeled robot equipped with infrared sensors is our actor. On it there are applied some strategies to obtain the reactive behaviour for unforeseeable obstacles avoidance. The reactive function is a mapping on the perceptual states set with values in the action set. A rule base perception action is obtained. The building of this is tackled using a genetic algorithm as well as by reinforcement learning.
To the mobile robot equipped with this ability the navigation skill is attached so that it is able to navigate from a start point to a goal one by dead reckoning by odometry. A control system based on two microcontrollers running in parallel and changing messages between each other is the hardware on which the control strategies are applied. Some results obtained by simulation of the developed strategies are presented.